Measures of narcissism often overlap with global self-esteem, risking that observed associations with outcomes may reflect associations of self-regard rather than actual narcissistic dispositions. The present study examined whether common narcissism instruments differ in their overlap with self-esteem and how this alters their associations with key personality domains. A sample of 337 participants completed multiple measures of narcissism, a global self-esteem measure as the control variable, and assessments of the Big Five, empathy, and aggression as personality correlates. Our results showed that overlap the measures of narcissism share with self-esteem varied considerably. Vulnerable scales showed the largest overlap and the greatest changes in correlations with the personality correlates after controlling for self-esteem. Grandiose and antagonistic measures were generally less affected, though noteworthy differences emerged between these instruments as well. We conclude that self-esteem overlap is a substantive but uneven measurement issue. Researchers cannot assume measures to be interchangeable. Our findings suggest that in order to isolate narcissistic dispositions from self-regard, researchers may need to select less affected instruments and/or report (additional) analyses controlling for self-esteem.